Download Construction of high-density bacterial colony arrays and

Construction of High-Density Bacterial
Colony Arrays and Patterns by the
Ink-Jet Method
Tao Xu,1 Sevastioni Petridou,1 Eric H. Lee,2 Elizabeth A. Roth,1
Narendra R. Vyavahare,1 James J. Hickman,1 Thomas Boland1
1
Department of Bioengineering, Clemson University, 502 Rhodes, Clemson,
South Carolina 29634; telephone: 864-656-7639; fax: 864-656-4466;
e-mail: tboland@ clemson.edu
2
Department of Microbiology, Clemson University, Clemson, South Carolina
Received 4 March 2003; accepted 20 May 2003
Published online 25 November 2003 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/bit.10768
Abstract: We have developed a method for fabricating bacterial colony arrays and complex patterns using commercially available ink-jet printers. Bacterial colony arrays with
a density of 100 colonies/cm2 were obtained by directly
ejecting Escherichia coli (E. coli) onto agar-coated substrates at a rapid arraying speed of 880 spots per second.
Adjusting the concentration of bacterial suspensions
allowed single colonies of viable bacteria to be obtained.
In addition, complex patterns of viable bacteria as well as
bacteria density gradients were constructed using desktop
printers controlled by a simple software program. B 2004
Wiley Periodicals, Inc.
Keywords: ink-jet printer; bacterial colony array; complex
pattern; cell density gradient; biosensor
INTRODUCTION
Ink-jet printing is a noncontact reprographic technique that
takes digital data from a computer representing an image or
character, and reproduces it onto a substrate using ink drops
(Mohebi and Evans, 2002). In recent years, the applications
of the technology have been successfully extended from
its traditional areas in electronics (Le, 1998) and microengineering (Mott et al., 1999) into bioengineering fields
such as genomics, biosensors, and drug screening (Blanchard et al., 1996; Hart et al., 1996; Lemmo et al., 1998).
Although biological molecules and structures are often
viewed as fragile, molecules such as DNA have been
directed onto glass by a Canon bubble jet printer to fabricate
high-density DNA microarrays without molecular degradation (Okamoto et al., 2000). Moreover, proteins such as
horseradish peroxidase have been deposited onto cellulose
paper to create active enzyme arrays for bioanalytical assays
(Roda et al., 2000). We have also shown recently that active
biosensors based on biotin-streptavidin linkages can be
deposited onto glass (Pardo et al., 2003). Although ink-jet
technology has proven feasible for a broad spectrum of
biological samples, until recently, most of its use has been
Correspondence to: Thomas Boland
B 2004 Wiley Periodicals, Inc.
limited to nonviable materials (Wilson and Boland, 2003).
Here we describe the use of modified ink-jet printers to
deliver viable E. coli DH5a cells onto a nutrient-covered
glass substrate to create high-density colony arrays and
organizationally complex patterns.
High-density colony arrays are routinely used for the
construction of genomic and expression libraries. These
require high-throughput screening of thousands of bacteria
to identify specific DNA sequences, to investigate gene
expression, and/or to search for differentially expressed
genes. Currently, fabrication of bacterial multi-colony
arrays is performed mainly by commercially established
robotic spotting systems, such as the Genentix QBot (http://
www.genetix.com) and the ORCA robot system (Copeland
and Lennon, 1994). These are expensive and complex
systems that employ a multi-pin contact-delivery mechanism to transfer bacteria onto a membrane filter. Moreover,
there is an increasing need for suitable methods to create
bacterial patterns to build bacterial biosensors for environmental monitoring, detection of toxicological contamination (Ptitsyn et al., 1997; Rettberg et al., 1999), and
applications in the area of public defense (Pancrazio et al.,
1999; Ringeisen et al., 2002). The method of printing
described here presents an easy and reliable way to create
complex patterns of viable bacterial cells.
MATERIALS AND METHODS
Thermal Ink-Jet Printer and Print Substrate
Two different commercial ink-jet printers (Hewlett-Packard DeskJet 550C and Canon Bubble Jet 2100) were
modified to accommodate delivery of bacteria to a precise
location on a certain substrate. Details of the modifications
are published elsewhere (Pardo et al., 2003). Print substrates were made from soy agar. 1.5 mL and 25 mL of
prewarm sterilized TrypticaseR Soy Agar solution (Becton
Dickinson & Co, Cockeysville, MD) was poured into
35 mm and 100 mm Petri dishes, respectively, containing
sterilized coverslips or a microscope slide. The solution
was allowed to cool to room temperature, at that point a
thin gel layer was formed on the substrates. A modified
Canon bubble jet printer was used to deliver the bacteria
onto the coated coverslips because of its narrow distance
between cartridge head and substrate-loading base. For
delivery of the bacteria onto the microscope slides, a
modified HP DeskJet printer was used, due to its wide
space between cartridge head and substrate-loading base.
of a cartoon tiger paw and black color gradient. Using
the modified Canon ink-jet printer, the 3107 cells/mL
suspension of bacteria was deposited onto an agar-coated
coverslip according to these specially designed patterns.
After printing, cells were incubated overnight. A digital
camera optically recorded these complex patterns of viable
E. coli.
RESULTS
Escherichia coli Colony Array Fabrication
Bacterial Strain and Suspension
Escherichia coli DH5a cells (Gibco-BRL, Life Technologies, Rockville, MD) were grown overnight at 37jC on a
TrypticaseR Soy Agar plate. Two loopfuls of organisms
(representing approximately two large colonies) were
transferred into a centrifuge tube containing 5 mL sterilized
water. This formed the original print suspension of bacteria.
The cell concentration in the E. coli solution was determined to be 3107 cells/mL by the standard plate count
method (Benson, 1998). This solution was diluted to different concentrations of bacterial suspensions for subsequent printing. The tubes containing bacterial suspensions
were forcefully shaken before printing, to break up clumps
and ensure good distribution of the bacteria. The movement
of the cartridge during printing allowed the cells to be maintained in suspension.
Fabrication of Colony Array
Microsoft PowerPoint software was used to edit a colony
array pattern with a 2 dots/mm density and a 0.13 pt weight.
A black ink-jet cartridge was emptied of its contents,
thoroughly washed, rinsed with a 100% ethanol solution,
and autoclaved water and dried in a sterilized hood before
being filled with 1 mL of a bacterial printing suspension.
This procedure proved to be an effective method for the
cleaning and sterilization of the cartridges. This was
determined by imprinting coverslips with sterilized waterfilled cartridges and subsequent absence of colonies on
agar-coated coverslips 2 days after incubation at 37jC.
Using the modified HP Desktop 550C printer, droplets of
the E. coli suspension were ejected onto the agar-coated
glass slides according to a predesigned colony array pattern.
Different dilutions of print suspensions (1:10, 1:100, and
1:1000 from the original print suspension) were also used.
All printed slides were incubated at 37jC overnight. The
E. coli colony arrays were then scanned using an Epson
Perfection 1640SU scanner, and the images were recorded
for further analysis.
Complex Pattern Printing
We also proceeded to print E. coli directly using some
irregular patterns and complex shapes, including the picture
30
The E. coli imprinted agar-coated slides, after incubation
for 20 h at 37jC, were placed onto a high-resolution
scanner, which provided an even light source for the
optimal recording of the imprinted colony arrays. Individual colonies had a circular shape with an approximate
diameter of 500 mm. In the following ‘‘designed dots’’
refers to the actuation of the nozzle according to software
instructions. The designed dots could grow into single
colonies, or overlapping colonies, which we refer to as
‘‘imprinted dots’’ hereafter. Designed dots not resulting in
imprinted dots were also possible at low concentration of
the bacterial print suspension.
Figure 1 shows the optical recording of the colony arrays
from different bacterial concentrations that were printed on
identically designed patterns. Few imprinted colonies are
observed on slides A and B, where the bacterial print
concentrations were 3104 and 3105 cells/mL, respectively. On slides C and D, where the bacterial print
concentrations were 3106 and 3107 cells/mL, respectively, the many bacterial colonies present matched the
designed array patterns. Overlapping colonies were observed on the slides with the highest bacterial concentrations. Furthermore, it appears that occasionally, during
the nozzle firing, satellite droplets formed, and resulted
Figure 1. Optical recordings by scanner of E. coli colony arrays created
using print suspensions of varying E. coli concentrations. Slide A: 3
104 cells/mL; slide B: 3105 cells/mL; slide C: 3106 cells/mL; slide
D: 3107 cells/mL. The insert shows a higher magnification of several
colonies on slide D, revealing an overlap and a satellite defect.
BIOTECHNOLOGY AND BIOENGINEERING, VOL. 85, NO. 1, JANUARY 5, 2004
in bacterial colonies growing outside the designed dots
(slide D).
To analyze the ability of our ink-jet printing method to
obtain optimal single-colony arrays we counted the relative
number of imprinted dots and isolated colonies on each
slide as a function of the concentration of the bacterial suspension (Fig. 2). Four rows (each containing 60 designed
dots) of bacterial colony arrays were analyzed. A bacterial
concentration of 3106 cells/mL resulted in single colony
arrays with the highest fidelity to the pattern.
Figure 3. Photograph of cartoon tiger paw generated by printing
viable E. coli cells, according to a predesigned pattern, onto an agarcoated coverslip.
Complex Patterns Printing
The cartoon tiger paw was designed as an example of
printing bacteria in an irregular shape. No obvious visual
pattern could be seen by the naked eye immediately after
printing the E. coli suspension onto the agar substrate.
However, after overnight incubation at 37jC, the pattern
became visible. The bacteria thus created the complex
image of a tiger paw (the unofficial Clemson University
logo). An optical micrograph of the tiger paw pattern obtained by direct printing of the E. coli suspension is depicted in Figure 3.
The black color gradient was designed to represent a
continuously changing pattern, where the darkness of color
should represent the higher density of imprinted bacteria.
The deposition of cells onto the soy agar surface with the
Canon printer following the grayscale pattern, showed a
bacterial density-gradient that changed from complete
coverage on the left side, to virtually no colonies on the
extreme right, hence closely matching the designed pattern.
The designed gradient pattern and the imprinted bacterial
density gradient (24 h after incubation at 37jC) are shown
in Figure 4.
Figure 2. Graphic representation of the relative number of imprinted
dots and isolated colonies, with respect to the density of the bacterial
suspension. ( ) Average percentage and Standard Deviation (n = 4) of
imprinted dots with respect to the total designed dots. (n) Average
percentage and Standard Deviation (n = 4) of isolated bacterial colonies
with respect to the total imprinted dots on the agar-coated slide.
DISCUSSION
Bacteria Printing and Fabrication of Colony Arrays
We have investigated the direct printing of bacteria on
agar-coated substrates using different ink-jet printers. Our
experiments provide evidence that viable E. coli cells can
be reliably delivered to target substrates thus making it
possible to produce high-density arrays and complex
patterns with a method that is fully automated and high
throughput. No nozzle blockage was observed and the
cartridges could be re-used many times, provided a cleaning protocol was employed.
A probable reason for absence of any reports of printing
bacteria by ink-jet printers is the prevailing thought that the
high temperate of the heating system and the high shear
stresses (up to 10 ms 1) (Okamoto et al., 2000) for ejected
samples in nozzles of ink-jet printers would damage or
even kill viable cells during their passage through the
nozzle. In Hewlett-Packard and Canon ink-jet printers a
similar print system is used, in which a heated plate in the
nozzle causes a vapor bubble to form and eject a droplet of
ink. The current pulse lasts a few microseconds and raises
the plate temperature as high as 300jC (Calvert, 2001). The
E. coli strain (DH5a) used was a competent cell strain
traditionally used for transformation and cloning. These
Figure 4. Cell density gradient of E. coli cells imprinted on an agarcoated coverslip, following a grayscale pattern. (A) Black color gradient
edited in Microsoft PowerPoint to provide patterns for cell density
gradient. (B) Scanner recording of E. coli density gradient created by direct
ink-jet printing of E. coli.
XU ET AL.: CONSTRUCTION OF HIGH-DENSITY BACTERIAL COLONY ARRAYS BY INK-JET METHOD
31
cells are more fragile and heat-sensitive than most other
E. coli strains. However, our results showed that although
the temperature of the heated plate can reach 200– 300jC,
the system still reliably delivered viable cells onto the agarcoated surfaces. One possible explanation for their survival
is that the limited amount of heat transferred to the ejected
droplets in the timeframe of the ejection causes only a
small temperature increase in the liquid and therefore little,
if any, damage occurs to the bacteria inside the droplets.
The optimal concentration of bacterial print suspension
was determined by both experimental and theoretical
means. Ink-jet printers can deliver as little as 8 pL/droplet,
or as much as 95 pL/droplet. The droplet sizes vary
according to the temperature gradient applied, the frequency
of the current pulse, and the sample viscosity (Allain et al.,
2000). The diameter of a single E. coli is approximately
2 Am (Harley et al., 2002) and the volume of the bacterium
is about 410 3 pL, which is about 1/2000 or 1/25,000 of
the volume of a droplet. Therefore, the concentration of
the printing suspension was carefully adjusted to achieve
printing of a single bacterium per drop. Based on the
specifications of each printer it was decided to use the value
of 95 pL/droplet to estimate the maximum concentration
of bacteria in the print suspension to ensure single colony
imprinting. Under these conditions the calculated maximum concentration of bacterial suspension needed was
estimated to be approximately 1.1107 cells/mL, but this
value was then fine-tuned by experimental observations.
In our experiments the optimum concentration was
found to be approximately 3106 cells/mL in the print
suspension to achieve approximately one bacterium per
droplet. When the concentration of the bacterial suspension
was 3107 cells/mL the probability that two or more
bacteria would be imprinted onto the same dot was higher
(Fig. 1, slide D). On the other hand, when lower concentrations were used, there was a much lower probability for
two or more bacteria to occupy the same site (Fig. 1, slides A,
B, C). Similarly, the results detailed in Figure 2 show that
using 3107 cells/mL concentrations resulted in 90%
isolated colonies. When a higher concentration of cells
was used, such as 1108 cells/mL, the percentage of isolated
colonies dropped to about 80% (data was not shown).
Moreover, no colony overlap was observed at concentrations
below 3107 cells/mL. In addition, some printed droplets
did not contain any bacteria, as shown by spots devoid of
bacterial colonies. The probability of the absence of bacterial
colonies increases as the concentration of the bacterial
suspension decreases. Therefore, there is a fine balance
between colony overlap and missing bacterial colonies. This
optimum concentration can also depend on the temperature
gradient applied, the frequency of the current pulse, the
sample viscosity, and the cell size. In summary, based on
these results the optimum concentration was determined to
be 3106 cells/mL, in good agreement with the calculated
volume of 1.1107 cells/mL.
The density of colony arrays created by the ink-jet print
method was about 100 dots/cm2 resulting in 48,400 colony
32
dots onto a 22 mm 22 mm positively charged nylon
membrane target filter. To construct colony arrays, the
conventional methods, such as the Genentix QBot or the
ORCA robot system, are based on a time-consuming multipin contact process that transfers bacteria onto a membrane
filter in ordered arrays. A gridding head with 96 or 384 pins
is used to dip into 96 or 384 well source plate to pick target
bacteria and then direct them onto the filter. Washing and
sterilization of the multi-pins must be performed before the
gridding head can be moved to a new 96/384-well source
plate. Conventional systems grid 3456 clones onto 812 cm
filters, far less than can be done using the ink-jet approach.
The gridding speed of conventional systems is about
100,000 spots/h (http://www.genetix.com), while the inkjet based systems, such as the one presented here, can
operate at 50,000 spots/min. Furthermore, the noncontact
mode of ink-jet printers eliminates the need of washing and
sterilization between depositions of bacteria. Additionally,
loading different bacteria strains into different color
cartridges provides a potential for creating mixed colony
arrays from different bacteria types.
Construction of Complex Patterns and Bacteria
Density Gradients
The capacity of desktop printers to orchestrate the printing
of bacteria into complex patterns is illustrated in Figures 3
and 4. In fact, the method is as simple as it is elegant. The
tiger paw pattern has an irregular shape, which is not
conveniently fabricated by conventional patterning methods. Competing methods, such as microcontact printing, for
example, would have difficulty creating complex arrangements of cells. Most methods use a two-step approach of
orchestrating cell patterns by employing masks to pattern
cell-adhesive substrates using self-assembled monolayers
or certain proteins (Bernard et al., 2000) to which the cells
are then exposed. The ink-jet approach is not only faster,
but it is also easy to automate. In addition, cell density
gradients are difficult, if not impossible to fabricate by the
two-step approach. Finally, with the ink-jet method the
creation of new patterns and modification of existing
patterns is simply accomplished by editing conventional
word-processing or graphics software.
CONCLUSION
We developed a noncontact ink-jet print method in which
viable bacteria can be delivered onto precisely targeted
positions that reproducibly generate isolated colony arrays,
as well as complex patterns. Our system provides a costand time-effective alternative to the traditional method for
creating arrays for bacterial colony screening and the
construction of genomic and expression libraries. In
addition, the ink-jet printer patterning technique, with its
simplicity and high throughput, may hold considerable
potential for the fabrication of bacterial biosensors. Moreover, the ink-jet printing could also be used to generate cell
BIOTECHNOLOGY AND BIOENGINEERING, VOL. 85, NO. 1, JANUARY 5, 2004
density gradients that may be of use in pharmacological
screening, where drug effectiveness and toxicity measurements require testing on varying densities of cells.
REFERENCES
Allain LR, Askari M, Stokes DL. 2001. Microarray sampling-platform
fabrication using bubble-jet technology for a biochip system. Fresenius
J Anal Chem 373:146 – 150.
Bernard A, Renault JP, Michel B, Bosshard HR, Delamarche E. 2000.
Microcontact printing of proteins. Adv Mater 12(14):1067 – 1070.
Benson HJ. 1998. Microbiological applications-Laboratory manual in
general microbiology, 7th ed. Boston, MA: WCB/McGraw-Hill.
p 89 – 94.
Blanchard AP, Kaiser RJ, Hood LE. 1996. High-density oligonucleotide
arrays. Biosens Bioelectron 11(6 – 7):687 – 690.
Calvert P. 2001. Inkjet printing for materials and devices. Chem Mat
13(10):3299 – 3305.
Copeland A, Lennon G. 1994. Rapid arrayed filter production using the
Orca Robot. Nature 369(6479):421 – 422.
Harley JP, Prescott LM, Klein DA. 2002. Microbiology. New York:
McGraw-Hill. p 500.
Hart AL, Turner APF, Hopcroft D. 1996. On the use of screen- and ink-jet
printing to produce amperometric enzyme electrodes for lactate.
Biosens Bioelectron 11(3):263 – 270.
Le HP. 1998. Progress and trends in ink-jet printing technology. J Imaging
Sci Techno 42(1):49 – 62.
Lemmo AV, Fisher JT, Geysen HM, Rose DJ. 1997. Characterization of an
inkjet chemical microdispenser for combinatorial library synthesis.
Anal Chem 69(4):543 – 551.
Lemmo AV, Rose DJ, Tisone TC. 1998. Inkjet dispensing technology:
Applications in drug discovery. Curr Opin Biotechnol 9(6):615 – 617.
Mohebi MM, Evans JRG. 2002. A drop-on-demand ink-jet printer for
combinatorial libraries and functionally graded ceramics. J Comb
Chem 4(4):267 – 274.
Mott M, Song JH, Evans JRG. 1999. Microengineering of ceramics by
direct ink-jet printing. J Amer Ceram Soc 82(7):1653 – 1658.
Okamoto T, Suzuki T, Yamamoto N. 2000. Microarray fabrication with
covalent attachment of DNA using bubble jet technology. Nat Biotechnol 18(4):438 – 441.
Pardo LF, Wilson Jr WC, Boland T. 2003. Characterization of patterned
self-assembled monolayers and protein arrays generated by the ink-jet
method. Langmuir 19:1462 – 1466.
Pancrazio JJ, Whelan JP, Borkholder DA, Ma W, Stenger DA. 1999. Development and application of cell-based biosensors. Ann Biomed Eng
27(6):697 – 711.
Ptitsyn LR, Horneck G, Komova O, Kozubek S, Krasavin EA, Bonev M,
Rettberg P. 1997. A biosensor for environmental genotoxin screening
based on an SOS lux assay in recombinant Escherichia coli cells. Appl
Environ Microbiol 63(11):4377 – 4384.
Rettberg P, Baumstark-Khan C, Bandel K, Ptitsyn LR, Horneck G. 1999.
Microscale application of the SOS-LUX-TEST as biosensor for
genotoxic agents. Anal Chim Acta 387(3):289 – 296.
Ringeisen BR, Chrisey DB, Pique A, Young HD, Modi R, Bucaro M, JonesMeehan J, Spargo BJ. 2002. Generation of mesoscopic patterns of
viable Escherichia coli by ambient laser transfer. Biomaterials 23(1):
161 – 166.
Roda A, Guardigli M, Russo C, Pasini P, Baraldini M. 2000. Protein
microdeposition using a conventional ink-jet printer. Biotechniques
28(3):492 – 496.
Wilson Jr WC, Boland T. 2003. Cell and organ printing 1: Protein and cell
printers. Anatom Rec 272A:497 – 502.
XU ET AL.: CONSTRUCTION OF HIGH-DENSITY BACTERIAL COLONY ARRAYS BY INK-JET METHOD
33